Haskell vs. Clean

[Philip Wadler]

Dear Ana,

(Sorry, Ana is your user name, but since you didn't sign your letter
I don't know what to call you.)

I'm forwarding your e-mail to Simon Peyton Jones, who is responsible
for the GHC implementation.  You imply that ease of installation is the
first step to widespread use, and I concur.  Myself, I use HBC from
Chalmers, which is much easier to install.

I'm surprised the Clean compiler performs so well on your benchmark.
In the published Pseudoknot benchmark (JFP 6(4):621--656, July 1996)
we find (along with data for twenty other compilers):

				precision	user + sys	memory
	Haskell (GHC)		single		 3.9 + 0.2	1.0 Mb
	Haskell (HBC)		single		12.1 + 0.1	3.0
	Clean			double		 5.1 + 0.8	2.5
	GNU CC			single		 2.4 + 0.1	0.3
	GNU CC			double		 2.7 + 0.1	0.3

Perhaps the Clean compiler has improved since then?  At any rate, I would
be interested to know how it achieves such good performance.

You also spoke of the various applications you make of Clean.  I hope
you will submit a description of these to the list of Real-World Applications
of Functional Programming (http://www.dcs.gla.ac.uk/fp/realworld/) and
write them up for the Journal of Functional Programming
(http://www.dcs.gla.ac.uk/jfp/).

Yours,  -- P

-----------------------------------------------------------------------
Philip Wadler                             wadler@research.bell-labs.com
Bell Laboratories                   http://www.cs.bell-labs.com/~wadler
Lucent Technologies                             office: +1 908 582 4004
600 Mountain Ave, room 2T-402                      fax: +1 908 582 5857
Murray Hill, NJ 07974-0636  USA                   home: +1 908 626 9252
-----------------------------------------------------------------------